Integration of plasma and CSF metabolomics with CSF proteomic reveals novel associations between lipid mediators and central nervous system vascular and energy metabolism

Authors: BORKOWSKI, KAMIL - University Of California, Davis; SEYFRIED, NICHOLAS - Emory University, School Of Medicine; ARNOLD, MATTHIAS - Duke University; LAH, JAMES - Emory University; LEVEY, ALLAN - Emory University; HALES, CHADWICK - Emory University; DAMMER, ERIC - Emory University, School Of Medicine; BLACH, COLETTE - Duke University; LOUIE, GREGORY - Duke University; KADDURAH-DAOUK, RIMA - Duke University; Newman, John

Submitted to: Scientific Reports
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/31/2023
Publication Date: 8/23/2023
Citation: Borkowski, K., Seyfried, N.T., Arnold, M., Lah, J.J., Levey, A.I., Hales, C.M., Dammer, E.B., Blach, C., Louie, G., Kaddurah-Daouk, R., Newman, J.W. 2023. Integration of plasma and CSF metabolomics with CSF proteomic reveals novel associations between lipid mediators and central nervous system vascular and energy metabolism. Scientific Reports. 13. Article 13752. https://doi.org/10.1038/s41598-023-39737-8.
DOI: https://doi.org/10.1038/s41598-023-39737-8

Interpretive Summary: Metabolomics and proteomics are modern tools allowing broad profiling of small molecules metabolites and proteins produced and present in body tissue and fluids that can provide insight into normal bodily functions. As proteins produce and integrate many biological signals, and metabolites often provide these signals, integrating metabolomic and proteomic data offers a unique opportunity to identify biochemical associations in metabolic disorders. Previously, we explored plasma and cerebrospinal fluid from individuals suffering from Alzheimer’s disease and independently identified both metabolomic and proteomics predictors of Alzheimer’s disease pathology and structural changes in the brain. In the current effort we sought to integrate these two data sets to explores associations between peripheral and central signaling lipids with the cerebrospinal fluid (CSF) protein markers of energy metabolism regulation, vasculature physiology, myelin/oligodendroglia, glia/inflammation, and synapses/neurons in the context of Alzheimer’s disease. We identified particularly strong associations between a suite of metabolites and CSF proteins involved in glycolysis, blood coagulation, vascular inflammation, and regulators of the extracellular matrix. The co-expression of genes linked to formation of these metabolites, or the associated proteins did not reveal this linkage. In summary, this metabolome-proteome integration provided new information regarding Alzheimer’s disease, linking both central and peripheral signaling lipids with AD-affected processes, while illustrating the utility of “omics” data integration to uncover associations beyond gene co-expression.

Technical Abstract: The integration of metabolomic and proteomic data provides a unique opportunity to identify biochemical associations in metabolic disorders. Using a combination of targeted metabolomics and proteomics, we previously identified a suite of plasma and cerebrospinal fluid metabolites lipid mediators as strong predictors of Alzheimer’s disease pathology, and five cerebrospinal fluid functional protein panels reflective of Alzheimer’s disease-related structural changes in the brain. The present effort explores associations between peripheral and central lipid mediators, including oxylipins, endocannabinoids and bile acids, with the cerebrospinal fluid protein panels markers of energy metabolism regulation, vasculature physiology, myelin/oligodendroglia, glia/inflammation, and synapses/neurons in the context of Alzheimer’s disease. We identified particularly strong associations between cytochrome p450/soluble epoxide hydrolase metabolites, bile acids and CSF proteins involved in glycolysis, blood coagulation and vascular inflammation and regulators of the extracellular matrix. We did not observe these metabolome-proteome associations at the gene-co-expression level in the central nervous system, supporting the utility of exploring pathway interactions in the distal region of the biochemical “omic” cascade. In summary, this metabolome-proteome integration effort provides new information regarding Alzheimer’s disease, linking both central and peripheral cytochrome p450/soluble epoxide hydrolase and bile acids metabolism with AD-affected processes, while illustrating the utility of “omics” data integration to uncover associations beyond gene co-expression.